A frontier is not a destination, and "promising" is not a property of an idea - it's a property of
whether the idea can be tested. So this is not a list of answers. It's a map of the open forks, sorted by
the one thing that separates a real frontier from a mirage: can the witness still reach it? A fork reality
can still collapse is the frontier. A fork that refuses to bet is a costume.
The whole sorting rule, in one line: a real frontier is a fork where reality can still
bite and hasn't yet. A mirage is a fork built so it can never be bitten. Beautiful math lives in both. Only
the first is science.
I · Live frontier — the witness can reach it now
These forks make distinguishable bets that current or near-future experiments can collapse. This is
where the real action is, because reality can still say no.
What is dark matter?
the longest open problem in physics
the forkA new particle (WIMP, axion, something lighter and weirder), or not a particle at all.
the testThree independent hunts: direct (xenon tanks underground, wait for a nucleus to recoil), indirect (telescopes watching for annihilation glow), production (collider missing-energy). Cross-checking by design.
statusThe detectors keep coming up empty. The favored WIMP is squeezed into smaller corners. The nulls are real information - the witness saying "not here," over and over.
OPEN · testable · the favored branch is dying
Modified gravity — maybe the bending is in the rule, not a hidden object
MOND · modified inertia · Verlinde's emergent gravity
the forkMaybe there's no missing matter - maybe the law that turns mass into motion is slightly wrong at galactic scales, and the "extra gravity" is a feature of the rule. The honest version of "the bending is an implementation artifact."
the testA real object can separate from the light it's near; a law-artifact can't. So: collide two galaxy clusters and see if the mass walks away from the gas.
statusWounded, not dead. It nails galaxy rotation curves better than dark matter - but fails the Bullet Cluster (where mass visibly parted from light) and the cosmic microwave background. The witness bit it, hard.
WOUNDED · half-falsified · wins small, loses big
The precision anomalies — the ASK gate, at civilizational scale
muon g−2 · W boson mass · the Hubble tension
the forkDoes a hyper-precise measurement disagree with the Standard Model in a way that survives every attempt to kill it? A crack, or a mistake?
the testRe-measure, re-calculate, demand independent replication. The Hubble tension is literally two independent measurements of the expansion rate (early-universe vs late) refusing to agree.
statusGenuinely unresolved and moving. Some anomalies have firmed up; others (muon g−2) got muddier as the theory side shifted. This is the field correctly in the ASK state - not guessing - while the tie resolves.
OPEN · the most likely shape of the next breakthrough
Why do neutrinos have mass?
the crack we already know is real
the forkThe Standard Model says neutrinos should be massless. They aren't (proven, Nobel 2015). So there's new physics here - a heavy partner particle? The reason matter beat antimatter?
the testHunt for neutrinoless double-beta decay (would prove the neutrino is its own antiparticle), measure the masses, watch the oscillations precisely.
statusA confirmed hole in a complete theory. We know the Standard Model is incomplete here, for certain - which makes it one of the most grounded frontiers there is.
OPEN · the incompleteness is proven
II · Dark frontier — the witness can't reach it yet
Real questions, possibly the deepest ones - but the place they'd be answered is past where any current
experiment can go. Not mirages (they'd make distinguishable predictions if we could test them) - just
beyond the witness's current reach. The honest word here is dark, not done.
Quantum gravity — unifying gravity with the rest
the great unfinished unification · lives below the Planck length
the forkCan gravity (curvature) and the quantum world be one theory? Strings, loop quantum gravity, something unborn?
the testWould require probing ~10⁻³⁵ m - where the energy to look closer collapses into a black hole and hides the answer. The microscope builds the wall.
statusThe biggest prize, the least reachable. Elaborate math on all sides, almost none of it currently distinguishable by experiment. Drift that can't yet be pruned - which is exactly why the field fights about whether it's progress or stalling.
DARK · real but currently un-witnessable
What was the Big Bang, and what (if anything) is "before"?
initial conditions · the white-hole question · the arrow of time
the forkWhy did the universe start in such low entropy? Is the Big Bang describable as a white hole? Where does time's arrow come from?
the testFaint - the CMB and gravitational-wave background carry whispers of the first instant, but the moment itself is behind a curtain no instrument has crossed.
statusPartly witnessable (the CMB is real data), mostly dark. The givens - the constants, the initial low entropy - are the hardcoded inputs the no-top theorem says you can read off but not derive.
DARK · the edge of the witnessable
III · Mirage — not a frontier, a costume
These feel like frontiers and explain nothing, because they're built so no observation could
ever contradict them. Beautiful, flexible, and un-prunable. Listed here not to mock but to mark the line - because
the difference between II and III is the whole discipline.
"We're in a simulation / it's the inside of a compiler"
the empty box, wearing sci-fi
the forkReality is computed; the anomalies are rendering artifacts.
the testNone. It's compatible with every possible observation, so it forbids none, so it predicts nothing. And it only relocates the mystery one floor up - the simulator's universe still has the same questions.
the rescueStrip the costume and there's a real idea inside: "maybe the bending is in the rule, not a hidden object." That idea is testable - it's modified gravity, up in section I, where it got bitten. Same instinct, made to bet. That's the move: take the un-prunable hunch and find the version that risks being wrong.
MIRAGE · un-prunable · but has a testable twin
So where is the next frontier? Not in any one of these answers - in the sorting. The live frontier
(dark matter's nature, the anomalies, the neutrino crack) is promising precisely because reality can still say no to
it. The dark frontier (quantum gravity, the first instant) is deeper and may be the real prize, but it's behind the
witness's reach, so honesty calls it dark, not solved. And the mirages aren't frontiers at all -
they're forks dressed up to be un-falsifiable, and the only thing to do with them is find their testable twin and
send that to the witness.
The promising thing isn't a theory. It's the method: drift forward to a fork, find the version that makes a
distinguishable bet, and let reality collapse it - over and over, in perpetuity, the frontier moving but never
closing because every resolved fork spawns its children. Dark matter is the cleanest case alive right now: 85% of
the matter in the universe, inferred for a century from how everything bends around it, never once touched - the
inferred void at cosmic scale - and a string of honest nulls slowly telling us the branch we confidently rode may be
the wrong one. That's not failure. That's the witness doing its slow, unglamorous, irreplaceable job: refusing to
confirm the favored guess until something real bites. The next frontier is wherever that refusal is still live.
Conceptual map, not computation - the gate here was factual accuracy, not a
Node check. The settled physics (Bullet Cluster, neutrino mass, the Planck wall) is solid; the anomaly statuses
(muon g−2, W mass, Hubble tension) move year to year and are presented as live-and-contested, not settled. For
the freshest state of any one fork, it's worth a live search - this is a snapshot, honestly dated.